Two Distinct N-Glycosylation Pathways Process the Haloferax volcaniiS-Layer Glycoprotein upon Changes in Environmental Salinity

Lina Kaminski,a Ziqiang Guan,b Sophie Yurist-Doutsch,a* Jerry Eichlera

Department of Life Sciences, Ben Gurion University of the Negev, Beersheva, Israela; Department of Biochemistry, Duke University Medical Center, Durham, NorthCarolina, USAb

* Present address: Sophie Yurist-Doutsch, Michael Smith Laboratories, University of British Columbia, Vancouver, BC, Canada.

ABSTRACT N-glycosylation in Archaea presents aspects of this posttranslational modification not seen in either Eukarya or Bac-teria. In the haloarchaeon Haloferax volcanii, the surface (S)-layer glycoprotein can be simultaneously modified by two differentN-glycans. Asn-13 and Asn-83 are modified by a pentasaccharide, whereas Asn-498 is modified by a tetrasaccharide of distinctcomposition, with N-glycosylation at this position being related to environmental conditions. Specifically, N-glycosylation ofAsn-498 is detected when cells are grown in the presence of 1.75 but not 3.4 M NaCl. While deletion of genes encoding compo-nents of the pentasaccharide assembly pathway had no effect on the biosynthesis of the tetrasaccharide bound to Asn-498, dele-tion of genes within the cluster spanning HVO_2046 to HVO_2061 interfered with the assembly and attachment of the Asn-498-linked tetrasaccharide. Transfer of the “low-salt” tetrasaccharide from the dolichol phosphate carrier upon which it is assembledto S-layer glycoprotein Asn-498 did not require AglB, the oligosaccharyltransferase responsible for pentasaccharide attachmentto Asn-13 and Asn-83. Finally, although biogenesis of the low-salt tetrasaccharide is barely discernible upon growth at the ele-vated salinity, this glycan was readily detected under such conditions in strains deleted of pentasaccharide biosynthesis pathwaygenes, indicative of cross talk between the two N-glycosylation pathways.

IMPORTANCE In the haloarchaeon Haloferax volcanii, originally from the Dead Sea, the pathway responsible for the assemblyand attachment of a pentasaccharide to the S-layer glycoprotein, a well-studied glycoprotein in this species, has been described.More recently, it was shown that in response to growth in low salinity, the same glycoprotein is modified by a novel tetrasaccha-ride. In the present study, numerous components of the pathway used to synthesize this “low-salt” tetrasaccharide are described.As such, this represents the first report of two N-glycosylation pathways able to simultaneously modify a single protein as a func-tion of environmental salinity. Moreover, and to the best of our knowledge, the ability to N-glycosylate the same protein withdifferent and unrelated glycans has not been observed in either Eukarya or Bacteria or indeed beyond the halophilic archaea, forwhich similar dual modification of the Halobacterium salinarum S-layer glycoprotein was reported.

Received 28 August 2013 Accepted 10 October 2013 Published 5 November 2013

Citation Kaminski L, Guan Z, Yurist-Doutsch S, Eichler J. 2013. Two distinct N-glycosylation pathways process the Haloferax volcanii S-layer glycoprotein upon changes inenvironmental salinity. mBio 4(6):e00716-13. doi:10.1128/mBio.00716-13.

Invited Editor Maria Hadjifrangiskou, Vanderbilt University School of Medicine Editor Scott Hultgren, Washington University School of Medicine

Copyright © 2013 Kaminski et al. This is an open-access article distributed under the terms of the Creative Commons Attribution-Noncommercial-ShareAlike 3.0 Unportedlicense, which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original author and source are credited.

Address correspondence to Jerry Eichler, jeichler@bgu.ac.il.

N-glycosylation, the covalent attachment of glycans to selectasparagine residues of target proteins, is a posttranslational

modification performed by members of all three domains of life(1–5). N-glycosylation in Archaea, although seemingly common(6) and involving a diversity of sugars not seen elsewhere (4, 5),nonetheless remains less well understood than the parallel eu-karyal and bacterial systems. Of late, however, a series of geneticand biochemical studies of Haloferax (Hfx.) volcanii (see reference5) and other species (7–12) have begun to provide insight into thearchaeal version of this universal protein-processing event.

In Hfx. volcanii, the surface (S)-layer glycoprotein, a well-studied glycoprotein and sole component of the protein shell sur-rounding cells of this species, is modified by a pentasaccharidecomprising a hexose, two hexuronic acids, a methyl ester of hexu-ronic acid, and a mannose (13, 14) via the actions of a series of Agl(archaeal glycosylation) proteins. Here, the glycosyltransferases

AglJ, AglG, AglI, and AlgE sequentially add the first four sugars ofthe pentasaccharide to a common dolichol phosphate (DolP) car-rier (15–18). Once the lipid-linked tetrasaccharide (and its pre-cursors) is translocated across the plasma membrane, the glycan isN-linked to the S-layer glycoprotein by the oligosaccharyltrans-ferase (OST) AglB (13). The final sugar of the N-linked pentasac-charide, mannose, is added to its own DolP carrier on the cyto-plasmic face of the membrane by AglD, translocated to face thecell exterior in a process involving AglR, and then transferred tothe protein-linked tetrasaccharide by AglS (17, 19–22). Finally,other Agl proteins, such as AglF, AglM, and AglP, also contributeto pentasaccharide assembly, serving various sugar-processingroles (14–23).

Hfx. volcanii, first isolated from the Dead Sea (24), requiresmolar concentrations of salt for survival. Recently, it was reportedthat N-glycosylation of the S-layer glycoprotein, in direct contact

RESEARCH ARTICLE

November/December 2013 Volume 4 Issue 6 e00716-13 ® mbio.asm.org 1

on February 29, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

with the hypersaline surroundings in which such cells exist,changes as a function of environmental salinity (25). S-layer gly-coproteins Asn-13 and Asn-83 bear the pentasaccharide describedabove in cells raised in either 1.75 M or 3.4 M NaCl-containinggrowth medium, although such modification is substantially re-duced upon growth at the lower salinity. More strikingly, S-layerglycoprotein Asn-498, whose position is not modified in cellsgrown at the higher salinity, is modified by a novel tetrasaccharidecomprising a sulfated hexose, two hexoses, and a rhamnose in cellsgrown at the lower salt concentration. The modulation of theS-layer glycoprotein N-glycosylation profile in response to chang-ing environmental conditions represents a novel role for this com-mon posttranslational modification.

To the best of our knowledge, the ability to N-glycosylate thesame protein with different and unrelated glycans has not beenobserved in either Eukarya or Bacteria or indeed beyond the halo-philic archaea, for which similar dual modification of the Halo-bacterium (Hbt.) salinarum S-layer glycoprotein was previouslyreported (26). Still, it is not known whether the Agl pathway,responsible for the assembly and attachment of the pentasac-charide bound at the Hfx. volcanii S-layer glycoprotein Asn-13and Asn-83 positions, is also involved in generating the “low-salt”tetrasaccharide bound to Asn-498 or whether a novel N-glycosylation pathway is responsible. In the present study, this wasaddressed.

RESULTSAssembly of the S-layer glycoprotein Asn-13- and Asn-83-linked pentasaccharide and the Asn-498-linked low-salt tetra-saccharide rely on different pathways. To determine whether thesame Agl pathway responsible for generating the pentasaccharideN-linked to S-layer glycoprotein Asn-13 and Asn-83 (13, 14) isalso responsible for generating the low-salt tetrasaccharide boundto Asn-498 (25), N-glycosylation at this position was consideredin Hfx. volcanii cells lacking AglI or AglE, glycosyltransferases in-volved in pentasaccharide assembly (15, 16), and grown in 1.75 MNaCl-containing (low-salt) medium. These mutants were selectedfor such analysis, as the shortened N-linked glycan decoratingAsn-13 and Asn-83 in each case is readily detected by mass spec-trometry.

To assess the contributions of AglI and AglE to Asn-498 glyco-sylation, liquid chromatography-electrospray ionization massspectrometry (LC-ESI MS) was employed. Initially, the LC-ESIMS profile of an S-layer glycoprotein-derived Asn-83-containingpeptide generated by trypsin and Glu-C protease treatment(65NQPLGTYDVDGSGSATTPNVTLLAPR90) from �aglI cellsgrown in low-salt medium was considered. Whereas no peptidemodified by the first three pentasaccharide sugars (m/z 1,597.675,calculated [M � 2H]2� mass) was detected (Fig. 1A, top, left), apeptide modified by the first two pentasaccharide sugars, a hexoseand a hexuronic acid, was observed ([M � 2H]2� peak at m/z1,491.70) (Fig. 1A, top, right). These results reflect the previouslyreported role of AglI in adding the third pentasaccharide sugar, ahexuronic acid, to disaccharide-charged DolP, from where it istransferred to S-layer glycoprotein Asn-83 (16). At the same time,when a total lipid extract of the �aglI cells grown in low-salt me-dium was examined, [M-2H]2� peaks at m/z 780.457 and 814.479,corresponding to C55 and C60 DolPs modified by a sulfatedhexose, two hexoses, and a rhamnose, namely, the complete low-salt tetrasaccharide (25), were observed (Fig. 1A, middle). The

same low-salt tetrasaccharide was also detected when an S-layerglycoprotein-derived Asn-498-containing peptide (497INGTASGANSVLVIFVD513; calculated [M � 2H]2� mass, 1,675.88 Da)from cells lacking AglI grown in low-salt medium was examinedby LC-ESI MS. The resulting profile included an m/z 1,195.03peak (Fig. 1A, bottom), as well as m/z 1,122.01, 1,040.97, and959.95 peaks (see Fig. S1 in the supplemental material, left col-umn). These values are in good agreement with, respectively, thepredicted [M � 2H]2� values of the Asn-498-containing peptidemodified by the complete low-salt tetrasaccharide (m/z 1,194.94)and precursors comprising the first three (m/z 1,121.94), the firsttwo (m/z 1,040.94), and the first (m/z 959.94) sugar componentsof this glycan (25). Likewise, in keeping with the role of AglE inadding the fourth pentasaccharide sugar, a hexuronic acid that issubsequently methylated by AglP (14, 15), the LC-ESI MS profileof proteolytic S-layer glycoprotein fragments from cells lackingAglE did not present a peak corresponding to the Asn-83-containing peptide modified by the first four pentasaccharide sug-ars (m/z 1,674.675, calculated [M � 2H]2� mass) (Fig. 1B, top,left). Instead, an [M � 2H]2� peak at m/z 1,579.72, correspondingto the same peptide modified by the first three sugars of the Asn-83-bound pentasaccharide was seen (Fig. 1B, top, right). LC-ESIMS analysis of a total lipid extract of the same cells revealed [M-2H]2� peaks at m/z 780.448 and 814.475, corresponding to C55

and C60 DolPs modified by the low-salt tetrasaccharide, respec-tively (Fig. 1B, middle). At the same time, the LC-ESI MS profile ofthe Asn-498-containing peptide from such cells presented [M �2H]2� peaks at m/z 1,195.04 (Fig. 1B, bottom), 1,122.03, and1,040.97 (Fig. S1, right column), respectively, corresponding tothe low-salt tetrasaccharide and its tri- and disaccharide precur-sors.

These observations thus point to the assembly of the low-salttetrasaccharide bound to S-layer glycoprotein Asn-498 as relyingon a biosynthetic pathway distinct from that used for assembly ofthe pentasaccharide that modifies the Asn-13 and Asn-83 posi-tions.

A novel N-glycosylation pathway modifies S-layer glycopro-tein Asn-498. Apart from aglD, all of the Hfx. volcanii genes in-volved in the assembly of the N-linked pentasaccharide decoratingS-layer glycoprotein Asn-13 and Asn-83 are organized into a sin-gle cluster spanning HVO_1517 (aglJ) to HVO_1531 (aglM) (16,27). To begin identifying components of the novel pathway in-volved in assembly of the low-salt tetrasaccharide, the Hfx. vol-canii genome sequence (28) was scanned for other clusters of openreading frames (ORFs) annotated as serving glycosylation-relatedroles. ORFs comprising the region spanning from HVO_2046 toHVO_2061 represents one such region. Table S1 in the supple-mental material summarizes the current annotations of these se-quences, as well as the predicted subcellular localization of thededuced products.

The participation of genes within the HVO_2046-2061 regionin generating the low-salt tetrasaccharide bound to Asn-498 wasnext considered by deleting each sequence within the cluster andassessing the effects of such deletion on Asn-498 glycosylation.The various deletion strains were generated by the “pop-in/pop-out” technique developed for use with Hfx. volcanii, in which thetrpA gene, encoding tryptophan synthase, is introduced into astrain auxotrophic for tryptophan in place of the gene being de-leted (29). In each case, deletion was confirmed (Fig. S2) and theeffect of each deletion was assessed by LC-ESI MS both at the level

Kaminski et al.

2 ® mbio.asm.org November/December 2013 Volume 4 Issue 6 e00716-13

on February 29, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

of the DolP carrier upon which the low-salt tetrasaccharide isassembled (25) and at the level of S-layer glycoprotein Asn-498.

In the case of �HVO_2058 cells, analysis of the LC-ESI MSprofile of a total lipid extract revealed C55 and C60 DolPs modifiedby the first three sugars of the low-salt tetrasaccharide ([M-2H]2�

ion peaks at m/z 707.415 and 741.443, respectively) (Fig. 2A, left)but not by the complete tetrasaccharide (Fig. 2A, right). Likewise,when the Asn-498-containing S-layer glycoprotein-derived pep-tide from the same cells grown in low-salt medium was analyzed,a [M � 2H]2� ion peak at m/z 1,122.03 was observed, correspond-ing to the peptide modified by the trisaccharide precursor of thelow-salt tetrasaccharide (Fig. 2B, upper). No peptide modified bythe complete tetrasaccharide was detected (Fig. 2B, lower).HVO_2058 is thus involved in the addition of rhamnose ontoDolP charged with the first three sugars of the low-salt tetrasac-charide. To confirm that the effect of HVO_2058 deletion on low-salt tetrasaccharide biogenesis was not due to a downstream effect,cells of the deletion strain were transformed to express a plasmid-derived version of the protein bearing a Clostridium thermocellumcellulose-binding domain (CBD) tag. The transformed strain wasable to assemble and attached the complete low-salt tetrasaccha-ride to S-layer glycoprotein Asn-498 (Fig. 2C). Finally, to gaininsight into HVO_2058 transcription as a function of growth me-dium salinity, quantitative PCR was performed. Relative to thelevel of HVO_2058 transcript measured in cells grown in 3.4 MNaCl-containing medium, 0.75-fold (�0.22-fold, standard devi-ation, n � 5) the transcript was seen in cells grown in 1.75 MNaCl-containing medium. As such, comparable levels of theHVO_2058 transcript are found at the different levels of salinity.

At the same time, C55 and C60 DolPs in �HVO_2048 cells weremodified by only the first two sugars of the low-salt tetrasaccha-ride ([M-2H]2� ion peaks at m/z 626.382 and 660.408) (Fig. S3A,left); no peaks corresponding to the same lipids bearing the firstthree sugars of the low-salt tetrasaccharide were detected(Fig. S3A, right). Analysis of the Asn-498-containing peptide fromthese cells revealed a [M � 2H]2� ion peak of m/z 1,040.97(Fig. S3B, upper), corresponding to the peptide modified by thefirst two sugars of the low-salt tetrasaccharide; no peptide modi-fied by the trisaccharide precursor of the Asn-498-linked low-salttetrasaccharide was observed (Fig. S3B, lower). As such, it can beconcluded that HVO_2048 is involved in the addition of the thirdsugar of the low-salt tetrasaccharide to DolP. To assess the effectsof growth medium salinity on HVO_2048 transcription, quanti-tative PCR was performed. Relative to the level of HVO_2048transcript measured in cells grown in 3.4 M NaCl-containing me-dium, 0.76-fold (�0.44-fold, standard deviation, n � 5) of thetranscript was seen in cells grown in 1.75 M NaCl-containing me-dium. It thus appears that the level of HVO_2048 transcriptiondoes not change substantially as a function of medium salinity.

Using the same experimental strategy, roles for HVO_2046,HVO_2049, HVO_2053, HVO_2055, HVO_2056, HVO_2057,

FIG 1 The Agl pathway is not involved in biogenesis of the low-salt tetrasa-ccharide. In �aglI and �aglE cells grown in low-salt medium, S-layer glyco-protein Asn-498 is modified by the low-salt tetrasaccharide. LC-ESI MS pro-files of S-layer glycoprotein-derived peptides containing Asn-83 (top) or Asn-498 (bottom) or of total lipid extracts (middle) from �aglI (A) and �aglE (B)cells. In each peptide-derived profile, arrows point to monoisotopic [M �2H]2� ion peaks corresponding to glycan-modified peptides or their expected

(Continued)

Figure Legend Continued

positions, while in the lipid-derived profiles, arrows point to monoisotopic[M-2H]2� ion peaks corresponding to low-salt tetrasaccharide-modifiedDolP. Adjacent to each arrow is a schematic depiction of the Asn- or DolP-bound glycan detected (or not), with the filled circles corresponding to hexose,the filled squares corresponding to hexuronic acid, and the open circle corre-sponding to rhamnose. In the middle panels, the arrows indicating �10 reflectmagnification of the ion peaks in the corresponding region of the profile.

Two Hfx. volcanii N-Glycosylation Pathways

November/December 2013 Volume 4 Issue 6 e00716-13 ® mbio.asm.org 3

on February 29, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

HVO_2059, HVO_2060, and HVO_2061 in Hfx. volcanii S-layerglycoprotein Asn-498 glycosylation were also revealed (Table 1).In the case of HVO_2047, deletion of the encoding gene did notreveal any effect on the low-salt tetrasaccharide at either the DolPor the S-layer glycoprotein level. As such, either HVO_2047 is notinvolved in N-glycosylation, as related to the low-salt tetrasaccha-

ride, or it mediates an effect not revealed by the massspectrometry-based analysis employed, as, for example, would bethe case for an epimerase.

Given the roles of products of genes in the HVO_2046-2061region in the biogenesis of the low-salt tetrasaccharide, thesewere renamed Agl5 (HVO_2053), Agl6 (HVO_2061), Agl7

FIG 2 Deletion of HVO_2058 affects DolP and Asn-498 glycosylation. (A) In cells lacking HVO_2058, C55 and C60 DolPs are modified by the first three sugarsof the complete low-salt tetrasaccharide (left) but not by the complete glycan (right). (B) The S-layer glycoprotein-derived Asn-498-containing peptide from�HVO_2058 cells is modified by the first three sugars of the low-salt tetrasaccharide (upper) but not by the complete glycan (lower). (C) In cells of the deletionstrain transformed with plasmid pWL-CBD2058, encoding a CBD-tagged version of HVO_2058, the S-layer glycoprotein-derived Asn-498-containing peptideis modified by the complete low-salt tetrasaccharide. The inset shows PCR amplification of CBD-HVO_2058 in the transformed strain but not in the deletionstrain. In each panel in the figure, the position of each detected or absent glycan-modified DolP or Asn-498-containing peptide is indicated, as is a schematicallydrawn depiction of the present or absent glycan. In the schematic drawings, the open circle represents rhamnose, while the filled circles represent hexose.

Kaminski et al.

4 ® mbio.asm.org November/December 2013 Volume 4 Issue 6 e00716-13

on February 29, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

(HVO_2046), Agl8 (HVO_2060), Agl9 (HVO_2048), Agl10(HVO_2049), Agl11 (HVO_2057), Agl12 (HVO_2059), Agl13(HVO_2056), Agl14 (HVO_2058), and Agl15 (HVO_2055), ac-cording to the nomenclature for archaeal N-glycosylation path-way components first proposed by Chaban et al. (7) and recentlyexpanded by Meyer et al. (10) (for further discussion, see reference30). Finally, the absence of Agl5-Agl15 did not compromiseAsn-13 and Asn-83 glycosylation (Fig. S4), reflecting the specificcontributions of these proteins to the assembly of the low-salttetrasaccharide.

The oligosaccharyltransferase AglB is not needed for Asn-498 glycosylation. Although Hfx. volcanii seemingly relies on twodifferent pathways for the assembly of the two N-linked glycansdecorating the S-layer glycoprotein, only one oligosaccharyltrans-ferase, AglB, has been identified in this organism (6, 31). In agree-ment with earlier studies (13), an S-layer glycoprotein-derivedAsn-83-containing peptide ([M � 2H]2� ion peak at m/z1,322.66) (Fig. 3A, upper) was not modified in cells deleted of aglB(Fig. 3A, lower). Next, to determine whether AglB also catalyzestransfer of the low-salt tetrasaccharide from DolP to Asn-498,�aglB cells were grown in low-salt medium and modification ofthe S-layer glycoprotein was considered by LC-ESI MS. Analysis ofthe S-layer glycoprotein-derived Asn-498-containing peptide re-vealed peaks corresponding to the peptide modified by the com-plete low-salt tetrasaccharide ([M � 2H]2� ion peak at m/z1,195.04) (Fig. 3B), as well as by the first ([M � 2H]2� ion peak atm/z 959.96) (Fig. S5, top), the first two ([M � 2H]2� ion peak atm/z 1,040.97) (Fig. S5, middle), and the first three ([M � 2H]2�

ion peak at m/z 1,122.02) (Fig. S5, bottom) low-salt tetrasaccha-ride sugars. It thus appears that AglB is not involved in the transferof the low-salt tetrasaccharide from its DolP carrier to S-layerglycoprotein Asn-498.

When grown in high salt, cells lacking glycosyltransferasesinvolved in Asn-13 and Asn-83 glycosylation decorate Asn-498with the low-salt tetrasaccharide. Finally, experiments were un-dertaken to assess whether interplay between the two Hfx. volcaniiN-glycosylation pathways exists. Specifically, it was tested whetherthe pathway responsible for biosynthesis of the low-salt tetrasac-charide could decorate Asn-498 in cells grown in medium con-taining 3.4 M NaCl when components of the pathway responsiblefor the assembly of the pentasaccharide N-linked to S-layer glyco-protein Asn-13 and Asn-83 were missing. Accordingly,N-glycosylation of Asn-498 was considered in cells deleted of aglIor aglE and grown in high-salt medium.

As previously reported (25), cells of the parent strain grown in

high-salt medium contained barely detectable amounts of C55 andC60 DolPs modified by the low-salt tetrasaccharide ([M-2H]2�

ion peaks detected at m/z 780.465 and 814.484, respectively)(Fig. 4A, inset). As expected, no low-salt tetrasaccharide waslinked to S-layer glycoprotein Asn-498 in parent strain cells raisedat the higher salinity (4A). When, however, the glycan-chargedpools of DolP from cells deleted of aglI or aglE and grown in 3.4 MNaCl were examined, considerably more low-salt tetrasaccharide-charged DolP was detected than in the parent strain, with theunrelated m/z 805.7 peak serving as an internal control in eachcase (upper panels, Fig. 4B and Fig. S6, respectively). At the sametime, analysis of the S-layer glycoprotein-derived peptides ob-tained from cells lacking AglI (Fig. 4B, bottom) or AglE (Fig. S6,bottom) and grown in high-salt medium revealed that low-salttetrasaccharide-modified Asn-498-containing peptide was readilydetected.

DISCUSSION

In Hfx. volcanii and in Hbt. salinarum, glycoproteins can be simul-taneously modified by two distinct N-linked glycans, each at-tached via a different linking sugar (25, 26). Such complexity inN-glycosylation has not been reported beyond these two archaealspecies. However, unlike Hbt. salinarum, where the pathway(s) ofN-glycosylation has not yet been characterized, much is known ofthe parallel process in Hfx. volcanii. In the present study, effortswere directed at determining whether the same pathway em-ployed for generating the pentasaccharide N-linked to Hfx. volca-nii S-layer glycoprotein Asn-13 and Asn-83 (13, 14) is also respon-sible for the assembly and attachment of the low-salttetrasaccharide N-linked to Asn-498 (25) of the same proteinwhen the cells are grown in low-salt medium.

Gene deletions, combined with mass spectrometric analysis ofglycan-charged DolP and the S-layer glycoprotein, revealed thatAgl proteins involved in the assembly of the Asn-13- and Asn-83-linked pentasaccharide do not participate in the biosynthesis ofthe low-salt tetrasaccharide attached to Asn-498. Instead, theproducts of a distinct set of genes mediate the assembly and at-tachment of this glycan. Given the substantial differences in thecompositions of the two N-linked glycans decorating the S-layerglycoprotein, it is not surprising that two distinct assembly path-ways are involved. On the other hand, the finding that the OSTAglB is not needed for low-salt tetrasaccharide attachment toS-layer glycoprotein Asn-498 is unexpected, as AglB is essential forN-glycosylation of Asn-13 and Asn-83 in cells grown in eitherhigh-salt or low-salt medium. In Hfx. volcanii, and indeed in Hbt.

TABLE 1 Genes involved in the biogenesis of the low-salt tetrasaccharidea

Deletion (gene designation) C55 or C60-DolP species observed S-layer glycoprotein Asn-498 glycosylation

�HVO_2046 (agl7) DolP-���Œ N498-��HVO_2048 (agl9) DolP-�SO3-� N498-�SO3- ��HVO_2049 (agl10) DolP-�SO3-�� N498-�SO3- ��

�HVO_2053 (agl5) DolP-� Not observed�HVO_2055 (agl15) DolP-�SO3-��Œ Not observed�HVO_2056 (agl13) DolP-�SO3-�� N498-�SO3- ��

�HVO_2057 (agl11) DolP-�SO3-�� N498-�SO3- ��

�HVO_2058 (agl14) DolP-�SO3-�� N498-�SO3- ��

�HVO_2059 (agl12) DolP-�SO3-�� N498-�SO3- ��

�HVO_2060 (agl8) DolP-�SO3-� N498-�SO3- ��HVO_2061 (agl6) DolP-� Not observeda Filled circles represent hexose, and the open circles represent rhamnose.

Two Hfx. volcanii N-Glycosylation Pathways

November/December 2013 Volume 4 Issue 6 e00716-13 ® mbio.asm.org 5

on February 29, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

salinarum, a single version of AglB is the only homologue of theeukaryal OST catalytic subunit, Stt3, or the bacterial OST, PglB,detected (6, 32). As such, a currently unidentified OST is involvedin the delivery of the low-salt tetrasaccharide (and its precursors)from DolP to S-layer glycoprotein Asn-498. The same may wellhold true in Hbt. salinarum, where one N-linked glycan is trans-ferred from a dolichol phosphate carrier and the second from adolichol pyrophosphate carrier (26, 33).

Based on the effects of deleting agl5-agl15 on DolP and S-layerglycoprotein glycosylation, a preliminary pathway for low-salt tet-rasaccharide biogenesis can be drawn (Fig. 5). In this workingmodel, Agl5 and Agl6 are assigned roles in adding the linkinghexose to DolP, while Agl7 contributes to the sulfation of thislipid-linked sugar. As such, the DolP-hexose seen in cells lackingAgl5 or Agl6 may correspond to this lipid charged with the firstsugar of the pentasaccharide transferred to Asn-13 and Asn-83, a

process that also occurs under low-salt conditions (25). More-over, because cells lacking Agl7 contained DolP charged with anonsulfated version of the low-salt tetrasaccharide, while no Asn-498-fused low-salt tetrasaccharide (or its di- or trisaccharide pre-cursors) were detected, sulfation of the DolP-bound hexose maybe needed for translocation of DolP charged with a more elaborateprecursor of the low-salt tetrasaccharide or the complete glycanacross the plasma membrane. Further work will be required todefine the precise actions of Agl5, Agl6, and Agl7, as well as theorder in which they act. Agl8 and Agl9 contribute to the additionof a hexose to disaccharide-charged DolP. Similarly, Agl10 toAgl14 are involved in the appearance of the final rhamnose sugarof the low-salt tetrasaccharide on the DolP carrier. In contrast,cells lacking Agl15 assemble the intact low-salt tetrasaccharide onDolP, yet no such glycan is observed on S-layer glycoprotein Asn-498. This effect is consistent with Agl15 acting as a flippase, medi-

FIG 3 AglB is not involved in S-layer glycoprotein Asn-498 glycosylation. (A) �aglB cells were grown in low-salt medium. LC-ESI MS analysis reveals theunmodified Asn-83-containing S-layer glycoprotein-derived peptide (upper) but not the same peptide modified by the normally bound pentasaccharide (lower).(B) LC-ESI MS analysis of the S-layer glycoprotein-derived Asn-498-containing peptide from cells lacking AglB reveals modification by the low-salt tetrasac-charide. In each panel, the position of the Asn-83 (A)- or Asn-498 (B)-containing peptide (detected or absent) is indicated, as is a schematically drawn depictionof the present or absent glycan. In the schematic drawings, the filled circles represent hexose, the filled squares represent hexuronic acid, and the open circlerepresents rhamnose.

Kaminski et al.

6 ® mbio.asm.org November/December 2013 Volume 4 Issue 6 e00716-13

on February 29, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

ating the translocation of low-salt tetrasaccharide-charged DolPacross the membrane. Accordingly, Agl15 shares 28% identity and51% similarity with AglR, recently proposed to serve as (or toassist) the flippase of DolP-mannose during the assembly of thepentasaccharide added to S-layer glycoprotein Asn-13 and Asn-83(22).

Presently, one can only speculate on why the Hfx. volcaniiS-layer glycoprotein is modified by the two distinct N-linked gly-cans under low-salt conditions but not at elevated salinity. WhileRNA encoding Agl5 to Agl15 was detected in cells grown at bothsalt levels, quantitative PCR-based studies will be necessary to de-termine whether comparable levels of these mRNAs are present atthe different salinities. A salt concentration-related conforma-tional change in the S-layer glycoprotein leading to exposure ofAsn-498 to the low-salt tetrasaccharide N-glycosylation machin-ery only at the lower salinity could be imagined. The fact that thisposition is not modified by the Asn-13/Asn-83-linked pentasac-charide when the cells are grown at higher salinity supports thishypothesis. However, the observation that barely detectableamounts of low-salt tetrasaccharide-bound DolP are seen underhigh-salt conditions argues that modification of Asn-498 is a ques-tion of the availability of this glycan for transfer to this S-layerglycoprotein residue. Why cells lacking different components ofthe N-linked pentasaccharide biosynthetic pathway decorate Asn-

FIG 4 Hfx. volcanii �aglI cells grown in 3.4 M NaCl-containing growth medium modify DolP and S-layer glycoprotein Asn-498 with the low-salt tetrasaccha-ride. (A) In parent strain cells, no Asn-498-containing fragment from the S-layer glycoprotein of parent strain cells grown in high-salt medium modified by thelow-salt tetrasaccharide was detected. (Inset) LC-ESI MS analysis of the low-salt tetrasaccharide-modified C55 and C60 dolichol phosphates in a total lipid extractof parent strain cells. (B) In �aglI cells, low-salt tetrasaccharide attached to Asn-498 is detected (lower). (Inset) LC-ESI MS analysis of a total lipid extract from�aglI cells reveals low-salt tetrasaccharide attached to C55 and C60 dolichol phosphates. The DolP-associated peaks shown in each upper panel correspond to[M-2H]2� ions, while Asn-498 peptide-associated peaks shown in each lower panel correspond to [M � 2H]2� ions. In each case, the lipid- or peptide-linkedlow-salt tetrasaccharide is schematically depicted at the indicated expected or observed position, with the open circle corresponding to rhamnose and the filledcircles corresponding to hexose.

FIG 5 Working model of the assembly of the low-salt tetrasaccharide deco-rating S-layer glycoprotein Asn-498 in Hfx. volcanii cells grown in 1.75 MNaCl. The sites of action of the novel Agl proteins identified in this study aredepicted. In the pathway, the vertical line corresponds to DolP, the full circlescorrespond to hexose, and the open circle corresponds to rhamnose. The Aglproteins listed act at the cytosolic face of the plasma membrane.

Two Hfx. volcanii N-Glycosylation Pathways

November/December 2013 Volume 4 Issue 6 e00716-13 ® mbio.asm.org 7

on February 29, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

498 with the low-salt tetrasaccharide at elevated salinity alsoawaits explanation.

In conclusion, this study has identified components of a sec-ond Hfx. volcanii N-glycosylation pathway involved in the post-translational modification of the S-layer glycoprotein. Indeed, ex-cept for the Hbt. salinarum S-layer glycoprotein, such dualN-glycosylation of the same protein has not been reported else-where. While it has been shown that the variant surface glycopro-tein of the eukaryotic parasite Trypanosoma brucei can present twodistinct N-linked glycans, the compositions of the two oligosac-charides are very similar (mannose5N-acetylglucosmine2 andmannose9N-acetylglucosmine2), both relying on the same linkingsugars. Moreover, both are added by isoforms of Stt3 (34, 35).Thus, the finding that AglB is not the OST of the novel Hfx. vol-canii N-glycosylation pathway is striking. As such, an enzyme notbelonging to the Stt3/PglB/AglB family of OSTs is apparently ca-pable of catalyzing the transfer of a lipid-linked glycan to a proteintarget. Future efforts will need to identify this protein, as well as toaddress other questions, including those related to the regulationof this novel N-glycosylation pathway, as well as its interplay withthe previously described Hfx. volcanii N-glycosylation pathwayinvolved in adding a pentasaccharide to select Asn residues of theS-layer glycoprotein.

MATERIALS AND METHODSStrains and growth conditions. Hfx. volcanii H53 parent strain cells (29)or the same cells deleted of aglB, aglE, or aglI (13, 15, 16, 29) were grownin medium containing 3.4 M NaCl (high salt) or 1.75 M NaCl (low salt),0.15 M MgSO4·7H2O, 1 mM MnCl2, 4 mM KCl, 3 mM CaCl2, 0.3%(wt/vol) yeast extract, 0.5% (wt/vol) tryptone, and 50 mM Tris-HCl,pH 7.2, at 42° C (36).

Quantitative PCR. Quantitative PCR was performed using Fast SYBRgreen master mix (Applied Biosystems, Foster City, CA). Primers (listedin Table S2) were designed using Primer Express 3.0 software (PerkinEl-mer Life Sciences) and employed at a final concentration of 500 nM. ForPCR amplification, 10-�l reaction mixtures were subjected to 40 reactioncycles in a StepOne real-time PCR system (Applied Biosystems). Primerefficiency was ascertained by drawing a standard curve based on 4-foldserial dilutions of cDNA when primers for HVO_2058 and _2048 wereused and 10-fold serial dilutions when primers for 16S rRNA were used.For expression analysis of HVO_2058 and _2048 in Hfx. volcanii H53 cellsgrown in medium containing 3.4 M or 1.75 M NaCl, 40 ng of cDNA wasused in 10-�l PCR amplifications. For measuring the levels of the 16SrRNA housekeeping gene, 5 pg of cDNA was used in a 10-�l reactionmixture. Relative mRNA levels were quantified using the standard2���CT formula.

Gene deletion. Deletion of Hfx. volcanii genes of interest was achievedusing a previously described “pop-in/pop-out” approach (29, 31). Toamplify approximately 500-bp-long regions flanking the coding se-quences of the various genes, appropriate primers directed against theupstream flanking regions and designed to introduce XhoI or KpnI andHindIII sites, and against the downstream flanking regions and designedto introduce BamHI and XbaI sites at the 5= and 3= of each fragment,respectively, were employed (Table S2). To confirm gene deletion at theDNA level, PCR amplification was performed using a primer directedagainst the upstream flanking region of the gene targeted for deletion,together with reverse primers against internal sequences of the gene inquestion or trpA, as appropriate (Table S2).

Complementation of the �HVO_2058 strain was achieved upontransformation with plasmid pWL-CBD2058, encoding HVO_2058 bear-ing an N-terminal C. thermocellum cellulose-binding domain tag. DNAencoding HVO_2058, bearing introduced NdeI and KpnI restriction sitesat the 5= and 3= ends, respectively, was ligated into plasmid pWL-CBD-

AglD (19) using T4 DNA ligase (Promega), following excision of theAglD-encoding region using NdeI and KpnI. To detect the presence ofHVO_2058 in the transformed cells, PCR amplification was performedusing appropriate primers (Table S2).

Mass spectrometry. LC-ESI tandem MS analysis of a total Hfx. volca-nii lipid extract was performed as described by Kaminski et al. (18), whileLC-ESI tandem MS analysis of the S-layer glycoprotein was performed asdescribed by Guan et al. (17).

SUPPLEMENTAL MATERIALSupplemental material for this article may be found at http://mbio.asm.org/lookup/suppl/doi:10.1128/mBio.00716-13/-/DCSupplemental.

Figure S1, PDF file, 0.5 MB.Figure S2, PDF file, 1.3 MB.Figure S3, PDF file, 0.3 MB.Figure S4, PDF file, 2.4 MB.Figure S5, PDF file, 0.3 MB.Figure S6, PDF file, 0.3 MB.Table S1, DOCX file, 0.1 MB.Table S2, DOCX file, 0.1 MB.

ACKNOWLEDGMENTS

J.E. is supported by the Israel Science Foundation (grant 8/11) and theU.S. Army Research Office (grant W911NF-11-1-520). The mass spec-trometry facility in the Department of Biochemistry of the Duke Univer-sity Medical Center and Z.G. are supported by Lipid Maps Large ScaleCollaborative grant GM-069338 from the NIH. L.K. is the recipient of aNegev-Zin Associates Scholarship.

REFERENCES1. Weerapana E, Imperiali B. 2006. Asparagine-linked protein

glycosylation: from eukaryotic to prokaryotic systems. Glycobiology 16:91R–101R.

2. Abu-Qarn M, Eichler J, Sharon N. 2008. Not just for Eukarya anymore:protein-glycosylation in Bacteria and Archaea. Curr. Opin. Struct. Biol.18:544 –550.

3. Dell A, Galadari A, Sastre F, Hitchen P. 2010. Similarities and differencesin the glycosylation mechanisms in prokaryotes and eukaryotes. Int. J.Microbiol. 2010:. PubMed148178.

4. Schwarz F, Aebi M. 2011. Mechanisms and principles of N-linked proteinglycosylation. Curr. Opin. Struct. Biol. 21:576 –582.

5. Eichler J. 2013. Extreme sweetness: protein glycosylation in Archaea. Nat.Rev. Microbiol. 11:151–156.

6. Kaminski L, Lurie-Weinberger MN, Allers T, Gophna U, Eichler J.2013. Phylogenetic- and genome-derived insight into the evolutionaryhistory of N-glycosylation in Archaea. Mol. Phylogenet. Evol. 67:327–339.

7. Chaban B, Voisin S, Kelly J, Logan SM, Jarrell KF. 2006. Identificationof genes involved in the biosynthesis and attachment of Methanococcusvoltae N-linked glycans: insight into N-linked glycosylation pathways inArchaea. Mol. Microbiol. 61:259 –268.

8. VanDyke DJ, Wu J, Logan SM, Kelly JF, Mizuno S, Aizawa S, Jarrell KF.2009. Identification of genes involved in the assembly and attachment of anovel flagellin N-linked tetrasaccharide important for motility in the ar-chaeon Methanococcus maripaludis. Mol. Microbiol. 72:633– 644.

9. Maita N, Nyirenda J, Igura M, Kamishikiryo J, Kohda D. 2010. Com-parative structural biology of eubacterial and archaeal oligosaccharyl-transferases. J. Biol. Chem. 285:4941– 4950.

10. Meyer BH, Zolghadr B, Peyfoon E, Pabst M, Panico M, Morris HR,Messner P, Schäffer C, Dell A, Albers SV. 2011. Sulfoquinovose syn-thase—an important enzyme in the N-glycosylation pathway of Sulfolobusacidocaldarius. Mol. Microbiol. 82:1150 –1163.

11. Jones GM, Wu J, Ding Y, Uchida K, Aizawa S, Robotham A, Logan SM,Kelly J., Jarrell KF. 2012. Identification of genes involved in the acet-amidino group modification of the flagellin N-linked glycan of Methano-coccus maripaludis. J. Bacteriol. 194:2693–2702.

12. Matsumoto S, Igura M, Nyirenda J, Matsumoto M, Yuzawa S, Noda N,Inagaki F, Kohda D. 2012. Crystal structure of the C-terminal globularidomain of oligosaccharyltransferase from Archaeoglobus fulgidus at 1.75-Åresolution. Biochemistry 51:4157– 4166.

13. Abu-Qarn M, Yurist-Doutsch S, Giordano A, Trauner A, Morris HR,

Kaminski et al.

8 ® mbio.asm.org November/December 2013 Volume 4 Issue 6 e00716-13

on February 29, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

Hitchen P, Medalia O, Dell A, Eichler J. 2007. Haloferax volcanii AglBand AglD are involved in N-glycosylation of the S-layer glycoprotein andproper assembly of the surface layer. J. Mol. Biol. 374:1224 –1236.

14. Magidovich H, Yurist-Doutsch S, Konrad Z, Ventura VV, Dell A,Hitchen PG, Eichler J. 2010. AglP is a S-adenosyl-L-methionine-dependent methyltransferase that participates in the N-glycosylationpathway of Haloferax volcanii. Mol. Microbiol. 76:190 –199.

15. Abu-Qarn M, Giordano A, Battaglia F, Trauner A, Hitchen PG, MorrisHR, Dell A, Eichler J. 2008. Identification of AglE, a second glycosyltrans-ferase involved in N-glycosylation of the Haloferax volcanii S-layer glyco-protein. J. Bacteriol. 190:3140 –3146.

16. Yurist-Doutsch S, Abu-Qarn M, Battaglia F, Morris HR, Hitchen PG,Dell A, Eichler J. 2008. aglF, aglG and aglI, novel members of a genecluster involved in the N-glycosylation of the Haloferax volcanii S-layerglycoprotein. Mol. Microbiol. 69:1234 –1245.

17. Guan Z, Naparstek S, Kaminski L, Konrad Z, Eichler J. 2010. Distinctglycan-charged phosphodolichol carriers are required for the assembly ofthe pentasaccharide N-linked to the Haloferax volcanii S-layer glycopro-tein. Mol. Microbiol. 78:1294 –1303.

18. Kaminski L, Abu-Qarn M, Guan Z, Naparstek S, Ventura VV, RaetzCR, Hitchen PG, Dell A, Eichler J. 2010. AglJ adds the first sugar of theN-linked pentasaccharide decorating the Haloferax volcanii S-layer glyco-protein. J. Bacteriol. 192:5572–5579.

19. Plavner N, Eichler J. 2008. Defining the topology of the N-glycosylationpathway in the halophilic archaeon Haloferax volcanii. J. Bacteriol. 190:8045– 8052.

20. Calo D, Guan Z, Naparstek S, Eichler J. 2011. Different routes to thesame ending: comparing the N-glycosylation processes of Haloferax vol-canii and Haloarcula marismortui, two halophilic archaea from the DeadSea. Mol. Microbiol. 81:1166 –1177.

21. Cohen-Rosenzweig C, Yurist-Doutsch S, Eichler J. 2012. AglS, a novelcomponent of the Haloferax volcanii N-glycosylation pathway, is adolichol phosphate-mannose mannosyltransferase. J. Bacteriol. 194:6909 – 6916.

22. Kaminski L, Guan Z, Abu-Qarn M, Konrad Z, Eichler J. 2012. AglR isrequired for addition of the final mannose residue of the N-linked glycandecorating the Haloferax volcanii S-layer glycoprotein. Biochim. Biophys.Acta 1820:1664 –1670.

23. Yurist-Doutsch S, Magidovich H, Ventura VV, Hitchen PG, Dell A,Eichler J. 2010. N-glycosylation in Archaea: on the coordinated actions ofHaloferax volcanii AglF and AglM. Mol. Microbiol. 75:1047–1058.

24. Mullakhanbhai MF, Larsen H. 1975. Halobacterium volcanii spec. nov.,

a Dead Sea halobacterium with a moderate salt requirement. Arch. Micro-biol. 104:207–214.

25. Guan Z, Naparstek S, Calo D, Eichler J. 2012. Protein glycosylation as anadaptive response in Archaea: growth at different salt concentrations leadsto alterations in Haloferax volcanii S-layer glycoprotein N-glycosylation.Environ. Microbiol. 14:743–753.

26. Lechner J, Wieland F. 1989. Structure and biosynthesis of prokaryoticglycoproteins. Annu. Rev. Biochem. 58:173–194.

27. Yurist-Doutsch S, Eichler J. 2009. Manual annotation, transcriptionalanalysis and protein expression studies reveal novel genes in the agl clusterresponsible for N-glycosylation in the halophilic archaeon Haloferax vol-canii. J. Bacteriol. 191:3068 –3075.

28. Hartman AL, Norais C, Badger JH, Delmas S, Haldenby S, Madupu R,Robinson J, Khouri H, Ren Q, Lowe TM, Maupin-Furlow J, Pohlsch-roder M, Daniels C, Pfeiffer F, Allers T, Eisen JA. 2010. The completegenome sequence of Haloferax volcanii DS2, a model archaeon. PLoS One5:e9605. doi:10.1371/journal.pone.0009605.

29. Allers T, Ngo HP, Mevarech M, Lloyd RG. 2004. Development ofadditional selectable markers for the halophilic archaeon Haloferax volca-nii based on the leuB and trpA genes. Appl. Environ. Microbiol. 70:943–953.

30. Eichler J, Jarrell KF, Albers S. 2013. A proposal for the naming ofN-glycosylation pathway components in Archaea. Glycobiology 23:620 – 621.

31. Abu-Qarn M, Eichler J. 2006. Protein N-glycosylation in Archaea: defin-ing Haloferax volcanii genes involved in S-layer glycoprotein glycosyla-tion. Mol. Microbiol. 61:511–525.

32. Magidovich H, Eichler J. 2009. Glycosyltransferases and oligosaccharyl-transferases in Archaea: putative components of the N-glycosylation path-way in the third domain of life. FEMS Microbiol. Lett. 300:122–130.

33. Wieland F, Dompert W, Bernhardt G, Sumper M. 1980. Halobacterialglycoprotein saccharides contain covalently linked sulphate. FEBS Lett.120:110 –114.

34. Izquierdo L, Schulz BL, Rodrigues JA, Güther ML, Procter JB, BartonGJ, Aebi M, Ferguson MA. 2009. Distinct donor and acceptor specificitiesof Trypanosoma brucei oligosaccharyltransferases. EMBO J. 28:2650 –2661.

35. Izquierdo L, Mehlert A, Ferguson MA. 2012. The lipid-linked oligosac-charide donor specificities of Trypanosoma brucei oligosaccharyltrans-ferases. Glycobiology 22:696 –703.

36. Mevarech M, Werczberger R. 1985. Genetic transfer in Halobacteriumvolcanii. J. Bacteriol. 162:461– 462.

Two Hfx. volcanii N-Glycosylation Pathways

November/December 2013 Volume 4 Issue 6 e00716-13 ® mbio.asm.org 9

on February 29, 2020 by guest

http://mbio.asm

.org/D

ownloaded from